Sepsis in neonates

Recognition of systemic neonatal sepsis

Signs are usually non-specific since other conditions cause similar clinical states (for example, cardiac or respiratory failure, metabolic disorders).

Often gut feeling of baby being 'not quite right'.

General features of neonatal sepsis

• Pallor, lethargy, jaundice
• Fever, hypothermia, temperature instability.

Note: One-third of confirmed sepsis cases are normothermic.

• Poor tolerance to handling
• Hypoglycaemia/hyperglycaemia
• Blood gas derangements (including acidosis and lactate accumulation).

Respiratory symptoms

• increased respiratory rate
• apnoea 
• grunting
• cyanosis.

Cardiovascular system symptoms

• tachycardia 
• bradycardic episodes
• poor perfusion
• hypotension.

Cutaneous symptoms

• petechiae 
• bruising
• bleeding from puncture sites.

Gastrointestinal (GIT) symptoms

• poor feeding
• vomiting
• abdominal distension
• feed intolerance
• bilious aspirates/vomits
• loose stools.

Central nervous system (CNS) symptoms

• lethargy
• irritability
• seizures.

Any baby who is unwell must be considered at risk of sepsis and appropriate antibiotics commenced as soon as possible after taking cultures.

Inability to obtain cultures should not delay administration of antibiotics.

Investigations for neonatal sepsis

General investigations

These include parameters important in assessment of general wellbeing of the infant including:

• blood gases
• serum electrolytes
• true blood glucose.

Infection related tests (septic workup)

Non-specific markers C-reactive protein (CRP):

• CRP rises approximately 12 hours after onset of sepsis and returns to normal within two to seven days of successful treatment.
• If the CRP remains elevated or rises after initial improvement care must be taken to look for possible collections, including endocarditis (particularly if 'long-lines' have been used) or fungal infection.
• CRP is raised in 85 per cent of episodes of confirmed sepsis with a specificity of 90 per cent.
• It can therefore be normal in cases of true sepsis and should be used in conjunction with clinical signs and culture results.

Full blood examination (FBE)

• The polymorphonucleocyte (PMN) count can be normal in one-third of cases of confirmed sepsis but can also be elevated in the absence of infection.
• Neutropenia in the face of confirmed sepsis can indicate that the baby is extremely unwell.
• A raised immature to total white cell ratio (I:T ratio > 0.3) is about 85 per cent sensitive and specific - particularly for early-onset sepsis.

Tests to identify the infective organism

For early-onset sepsis

• Blood culture (if possible). Do not delay treatment if you cannot obtain cultures in an unwell baby
• Lumbar puncture (LP) should be performed where the 'index of suspicion' of meningitis is high (abnormal conscious state or seizures).
• LP may need to be delayed until after the infant's condition has stabilised sufficiently to tolerate the procedure and abnormalities of coagulation status have been controlled.
• If the initial blood culture is positive. LP must be performed to exclude meningitis since the presence of meningitis alters the length of antibiotic treatment as well as prognosis. A lumbar puncture may still be useful within four hours of commencing antibiotics as growth may still occur.
• There is little to be gained from performing urine aspiration for culture as haematogenous spread is the mechanism behind positive urine cultures in the first few days of life.

For late-onset sepsis

• Blood cultures (if possible). Do not delay treatment if you cannot obtain cultures in an unwell baby.
• SPA specimen of urine should be obtained as a primary UTI is not uncommon as a cause of sepsis after five days of age.
• The role of LP in late onset sepsis is controversial and depends on the clinical setting. A lumbar puncture may still be useful within four hours of commencing antibiotics as growth may still occur.

Non-NICU infants suspected of being septic

• LP should be performed to exclude CNS infection.
• If there is a high clinical index of CNS infection, appropriate treatment should be instituted early even if the LP is delayed until the baby is stable enough to tolerate the procedure.

Infants in NICU

• The role of LP is limited since the commonest organism causing sepsis is the coagulase-negative Staph (CONS).
• CONS rarely cause CNS infection unless a Ventriculo peritoneal shunt is present.
• LP when CONS is isolated from blood culture is reserved for infants who are not following the expected clinical course despite appropriate antibiotics.
• LP is performed when the infant's condition is suggestive of meningitis or blood culture identifies an organism other than CONS.
• Endotracheal Tube (ETT) cultures and skin swabs are of limited value for babies in Level 6 Neonatal units.
• Their value is as a guide to the profile and sensitivity of organisms in the nursery, particularly Staphylococcus aureus.

Place of care

• The baby with confirmed sepsis should be managed in a level 3-5 Neonatal unit where they can be observed closely.
• However in some cases where antibiotics are commenced whilst sepsis is being ruled out (for example, brief unexplained respiratory distress or the GBS positive mother with inadequate intrapartum antibiotic prophylaxis) the baby is clinically well and the septic markers are benign.
• In these cases it may be appropriate for the baby to be managed in the postnatal ward so as to keep mother and baby together.
• In such cases it may be possible for the IV cannula to be inserted and the IV antibiotics given in the postnatal ward.

However, the following caveats must apply:

• Nursing staff caring for the baby must be competent to do so (as part of their employment be rostered from time to time in the neonatal unit).
• The baby is medically cleared by the paediatrician to be managed in the postnatal ward on a case-by-case basis.
• The baby remains under the care of the paediatrician(s).
• The baby has regular observations of temperature, pulse rate and respiratory rate with IV cannula flushes.  Any abnormalities of these parameters must result in readmission to the neonatal unit.

Where there is Hospital in the Home (HITH) facilities, consideration may be given to completing the final dose(s) of antibiotics at home.

General measures

• In addition to the administration of antibiotics, great attention to supportive care is needed.
• Antibiotics should be considered as only part of the management of a septic neonate.

General

• thermal care
• incubator nursing
• phototherapy if warranted
• monitoring of oxygen saturation, heart rate and blood pressure

Respiratory

• support for apnoea
• hypoxia
• hypercapnoea
• respiratory distress

Cardiovascular

• plasma volume expanders (normal saline - 10-20 mL/kg initially)
• inotrope support is often needed and transfer to a Level 5-6 neonatal unit may be required
• correction of fluid, electrolyte, glucose and haematological derangements (including blood, platelets and clotting factors)
• an unstable infant usually needs enteral feedings withheld

Antibiotic choice

• Given the usual causative organisms the following regimes are recommended initially.
• Antibiotic choice can then be rationalised on the basis of culture results and clinical course.

Early-onset sepsis

Benzylpenicillin

Dose - 60 mg/kg/dose IV 12-hourly.
120 mg/kg/dose 12-hourly (if meningitis suspected).

Gentamicin

Dose - 5mg/kg/dose IV/IM.
Frequency - 36-hourly if >= 1200 g, 48 hourly if < 1200 g.

Notes:

• Babies beyond 1 week of age may need more frequent dosing.
• Both Benzylpenicillin and Gentamicin can be given IM if IV access is not possible but this is not preferable for ongoing care: an umbilical venous catheter should be considered.
• Gentamicin is physically incompatible with beta-lactam antibiotics (such as benzylpenicillin) - it is preferable to separate administration by 1 hour. If it is not possible to separate doses, ensure IV lines are adequately flushed with 0.9% sodium chloride before and after administration of these antibiotics.
• If history or clinical appearance suggests the possibility of Listeria, amoxycillin 50 mg/kg IV 12-hourly can be used instead of benzylpenicillin (although data indicating that this is superior is lacking).

Treatment of meningitis

Until sensitivities are known:

• Cefotaxime - 50 mg/kg/dose 12-hourly for preterm babies or term babies in the first week of life, eight-hourly after that time
• Amoxycillin - 50 mg/kg/dose 12-hourly for preterm babies or term babies in the first week of life, eight-hourly after that time

Late-onset sepsis

• Flucloxacillin and gentamicin are the usual first choice antibiotics except in suspected septic shock due to Gram-negative organism use vancomycin, gentamicin +/- vancomycin
• Meningitis use amoxicillin and cefotaxime
• Necrotising enterocolitis use amoxicillin, gentamicin, metronidazole

Vancomycin dose

The dose chosen needs to be guided by the clinical picture and age of patient, and adjusted according to trough levels.

• IV: Loading dose of 15 mg/kg then
• Preterm: 10 mg/kg/dose 24-hourly
  Term: Week 1 of life: 10 mg/kg/dose 12-hourly
  Week 2-4 of life: 10 mg/kg/dose eight-hourly
• Severe infections: 15 mg/kg/dose

Note: There is limited evidence behind dosing in preterm infants, and other centres may use alternative dosing protocols based on weight.

Flucloxacillin 25 mg/kg/dose 12-hourly for preterm babies or term babies in the first week of life, six to eight-hourly after that time.

Note: Doses of antibiotics need to be adjusted for age of the baby and on the basis of levels in the case of gentamicin and vancomycin. Refer to local drug protocols for monitoring guidelines.

An aminoglycoside other than gentamicin may be used in some hospitals at times depending on the profile of prevalent organisms.

When to stop antibiotics

Duration of antibiotic treatment depends upon the clinical condition of the infant and the organism identified on culture.

• Where the likelihood of infection is low, with a baby in good condition and infective indices negative, antibiotics can be ceased if cultures are negative after 48 hours.
• Sepsis strongly suspected, despite negative blood culture at 48 hours. It is advisable to repeat blood culture and continue antibiotics for at least five days providing infective indices have normalised. Another approach is to continue antibiotics for 48 hours after indices are normal.
• Proven Gram-negative bacteraemia, with clear CSF, treat for 10 days, antibiotics can be rationalised in the face of culture and sensitivities.
• Proven GBS bacteraemia, with clear CSF, 10 days treatment should be sufficient.
• Meningitis, treat for 14 days for GBS and 21 days for Gram-negative organisms. In some units, 48-hourly LPs are performed in cases of Gram-negative meningitis, with treatment continuing for 14 days after the first negative culture - in practice this usually equates with a 21-day treatment course.
• Urinary tract infection (UTI):
  • Treat with IV antibiotics for at least five days; a total of 10 days treatment is needed.
  • The infant can be managed with appropriate oral antibiotics for the latter half of the treatment course if clinical condition is satisfactory.
  • Ongoing prophylactic antibiotics will be needed until renal investigations (ultrasound and/or MCU) are completed.

The GBS-positive mother:

• At birth approximately 15 per cent of women are colonised with GBS.
• Up to 70 per cent of infants born to colonised women are themselves colonised.
• Infection occurs in 1 per cent of colonised infants.
• 75 per cent of early onset GBS disease in neonates occurs in term babies.
• The incidence of GBS disease varies, with the rate being three per 1,000 live births in the USA, compared to 0.3 per 1,000 in Australia and the UK.
• The risk is three times higher in the Aboriginal community.
• Screening for GBS remains the subject of heated debate, but it is known that intrapartum administration of antibiotics (penicillin or amoxycillin) reduces neonatal colonisation by 90 per cent and early onset GBS disease by 90 per cent.
  The CDC in the USA recommends that all women be screened with anorectal and vaginal swabs at 35-37 weeks' gestation.

Intrapartum antibiotics are given according to the following strategies.

If screening is performed administer to:

• GBS colonised women
• non-colonised women with risk factors present.

If screening is not performed administer to women with these risk factors:

• preterm onset of labour (< 37 weeks)
• ROM for > 18 hours
• maternal fever (> 38C)
• previous baby with invasive GBS disease
• GBS bacteriuria this pregnancy.

Use of the CDC guidelines is estimated to result in around 27 per cent of women receiving antibiotics, with an associated reduction in early onset GBS disease of around 85 per cent.

The disadvantages of such an approach are the risk of maternal complications (anaphylaxis) and the cost (GBS rates of > 0.5 per 1,000 live births are needed to justify such an approach on a cost-effectiveness basis).

Prolonged rupture of membranes (PROM)

• The majority of women will come into labour within 24 hours of rupture of the membranes; however, this may be delayed in up to 4 per cent of cases.
• PROM for greater than 18 hours may lead to increased risk of infection in mother and baby, particularly if the mother is GBS positive or undergoes repeated vaginal examinations.
• In practice, the risk is greatest for preterm infants, but 75 per cent of early onset GBS sepsis occurs in term babies.
• Since there is a lack of evidence from trials available there is debate as to the role of prophylactic antibiotics in PROM.
• Obstetric staff will need to consider signs of possible maternal sepsis, as well as risk factors such as GBS colonisation in deciding to administer antenatal antibiotics.
• Babies born with a background of PROM need to be viewed as potentially at risk of sepsis.

Preterm infants

Preterm infants, particularly those < 35 weeks, should be screened for sepsis and treated with IV antibiotics until infection in the baby has been excluded.

Term infants

• If there are no risk factors, apart from the PROM, the infant is usually observed closely and treated only if symptoms develop.
• If there is a risk factor present in addition to PROM, such as GBS positive mother, maternal intrapartum fever or suspected chorioamnionitis that infant should be closely observed for potential sepsis in hospital (heart rate, respiratory rate, temperature before feeds ) for at least 24 hours even if completely asymptomatic.
• Any symptomatic baby needs septic evaluation performed and treatment for infection regardless of the presence or absence of risk factors.

Fungal sepsis

• Generally seen in VLBW infants in Level 6 neonatal units.
• Risk factors include multiple courses of IV antibiotics, presence of central lines and extensive areas of skin breakdown.
• Consideration of fungal sepsis is particularly necessary when such infants deteriorate while receiving antibiotics.
• Empirical treatment with amphotericin until cultures are reported as clear for fungal organisms is appropriate.
• Suprapubic aspiration (SPA) of urine must be performed prior to starting Amphotericin as bag specimens will often be contaminated with Candida colonising the skin.
• If fungal sepsis is confirmed, then the addition of a further antifungal (for example, fluconazole 12 mg/kg 72-hourly (< 14 days) and 48-hourly (15-28 days), may be useful.
• Duration of treatment depends upon the site of infection but generally ranges from three to six weeks.
• Ultrasound of the kidneys and formal fundoscopy should be performed.

Areas of uncertainty in clinical practice

Role of antigen tests for GBS is controversial

• Urine specimens for GBS antigen can be positive when babies are colonised, even when a SPA specimen is taken.
• If a bag specimen is used, then contamination with skin GBS colonisation will result in a positive test.
  Antigen tests are more sensitive and specific for CSF specimens but cannot be relied upon to exclude infection.
• Antigen testing results need to be viewed from the point of view of adding supplementary evidence of possible infection but cannot be relied upon to prove or disprove GBS infection, and are thus of limited value.
• Similar limitations exist in testing for other bacterial antigens.

Antifungal prophylaxis

• A recent Cochrane review failed to demonstrate a reduction in fungal colonisation among patients receiving prophylactic oral nystatin compared with placebo although use still occurs in babies < 1,000 g.
• All patients in these trials were immunocompromised but beyond the neonatal period.
• A RCT of intravenous fluconazole compared to placebo during the first six weeks of life in 100 infants of less than 1,000 g birthweight showed a reduction in fungal colonisation and invasive fungal infection.

Treatment with granulocyte colony stimulating factor (G-CSF)

• G-CSF has been shown to increase PMN counts in VLBW babies, but the effect on sepsis reduction or mortality from sepsis has not been demonstrated.

Intravenous immunoglobulin (IVIG)

• Studies involving IVIG show a possible improvement in mortality in babies given IVIG as part of the treatment of sepsis.
• However, larger trials are needed to examine the role of IVIG in neonates with sepsis.

Other ancillary treatments that have been used include exchange transfusion and neutrophil transfusions, but insufficient data is available to recommend their use.

• Hospital in the Home

Further reading

• Isaacs D, Moxon ER. Handbook of Neonatal Infections - a practical guide. WB Saunders, London. 1999.
• Remington JS, Klein JO. Infectious Diseases of the Fetus and Newborn Infant 5Th Ed. WB Saunders, Philadelphia. 2000.
• CDC (Center for Diseases Control). Prevention of perinatal group B streptococcal diseases: a public health perspective. MMWR 1996: 45(RR-7).
• Smaill, F. Intrapartum antibiotics for Group B streptococcal colonisation. Cochrane Pregnancy and Childbirth Group. The Cochrane Library.
• Flenady, V. King, J. Antibiotics for prelabour rupture of membranes at or near term. [Protocol] Cochrane Pregnancy and Childbirth Group. The Cochrane Library. 
• Vancomycin